The design of selectively-activated anti-cancer prodrugs for use in antibody-directed and gene-directed enzyme-prodrug therapies.

Systemic anti-proliferative agents (cytotoxins) have been the most successful single design concept for anti-cancer drugs. However, they have inherent limitations (they target dividing cells rather than cancer cells) which limit their clinical efficacy, especially toward the more slowly-growing solid tumours. New concepts are required to improve the selectivity of their killing of tumour cells. One possibility is the use of prodrugs which can be activated selectively in tumour tissue. Several potential mechanisms for this are being explored, including tumour hypoxia, low extracellular pH, therapeutic radiation and tumour-specific endogenous or exogenous enzymes. In the last approach the exogenous enzyme can be delivered by attachment to monoclonal antibodies (ADEPT) or as DNA constructs containing the corresponding gene (GDEPT). A limitation of both approaches is that only a small proportion of the tumour cells become activation-competent, but this can be substantially overcome by the design of appropriate prodrugs capable of killing activation-incompetent cells via a bystander effect. We have proposed a modular approach to prodrug design in which a trigger unit determines tumour selectivity and an effector unit achieves the desired level of killing of cells when the trigger is activated. For ADEPT and GDEPT prodrugs the primary requirement of the trigger is efficient and selective activation by the appropriate enzyme; the released effector must be a potent, diffusible cytotoxin which fully exploits the small proportion of cells capable of activating the prodrug. A wide variety of chemistries has been used, but many of the existing effectors do not have all of these properties. We report work on two types of cytotoxin derived from very potent anti-tumour antibiotics (enediynes and amino-seco-cyclopropylindolines) as effectors in prodrugs for ADEPT and GDEPT applications.